Alanne, K., Salo, A., Saari, A. and Gustafsson, S.-I. 2006. Multi-criteriaevaluation of residential energy supply systems. Energy and Buildings, acceptedfor publication.

© 2006 by authors and © 2006 Elsevier Science

Preprinted with permission.

Multi-Criteria Evaluation of Residential Energy Supply Systems

Kari Alanne a,*, Ahti Salo b, Arto Saari c, Stig-Inge Gustafsson d

a Dept. of Mechanical Engineering, Laboratory of Heating, Ventilating and Air-Conditioning,

Helsinki University of Technology, P.O.Box 1100, 02015 TKK, Finland b Dept. of Engineering Physics and Mathematics,

Systems Analysis Laboratory,

Helsinki University of Technology, P.O.Box 1100, 02015 TKK, Finland c Dept. of Civil and Environmental Engineering,

Laboratory of Construction Economics and Management,

Helsinki University of Technology, P.O.Box 2100, 02015 TKK, Finland d Dept. of Mechanical Engineering,

Division of Energy Systems,

University of Linköping, 58183 Linköping, Sweden

Abstract

In this paper, we consider the selection of a residential energy supply system as a

multi-criteria decision-making problem, which involves both financial and

environmental issues. Specifically, we compare micro-CHP (micro-cogeneration)

heating with traditional heating systems through an evaluation that accounts for (i)

the decision-makers’ subjective preferences, (ii) uncertainties in the performance of

micro-CHP heating systems (which are partly caused by the lack of long-term

operational experiences), (iii) the context-dependency of life-cycle costs and

environmental burdens of heating systems. Motivated by these considerations, we

employ the PAIRS multi-criteria decision-making methodology that captures

incomplete information by way of interval-valued parameters and provides support

for sensitivity analyses, too. Our comparative analysis of alternative heating systems

suggests that micro-CHP is a reasonable alternative to traditional systems,

particularly from the environmental point of view.

Keywords: multi-criteria evaluation, micro-cogeneration, micro-CHP, residential

buildings

* Corresponding author: Tel.:+358-50-4680892; e-mail: kalanne@iki.fi

2

1 Introduction

The introduction of sustainable energy systems is strongly associated to what choices

real estate owners make among different heating options [1]. Earlier on, heating

systems have been selected largely on the basis of capital costs and energy costs.

Nowadays, such decisions are typically made on the basis of life-cycle costs;

however, as environmental issues become increasingly important, it is necessary to

consider the selection of technological alternatives a multi-criteria problem. Here,

one of the challenges is that it may be difficult to obtain complete information about

the decision-makers’ preferences. Furthermore, there are no long-term experiences

from new technological systems, wherefore the performance of alternative systems

with regard to the evaluation criteria involves uncertainties.

Although real estate owners tend to prefer “soft” and “ecological” values, only few

multi-criteria comparisons of residential heating systems have thus far been

published: for example, Riihimäki et al. [2] and Huovila et al. [3] note that in Finland,

good environmental value, good comfort, and safety are among the most important

criteria. One possible reason for the apparent paucity of multi-criteria studies is that

conventional multi-criteria approaches do not address uncertainties explicitly. A

further reason may be that the costs have often remained the overriding criterion in

the final decision.

More broadly seen, however, multi-criteria optimization methods have found uses in

the analysis of residential buildings. Li et al. [4] define a “green heating system” and

employ a multi-criteria optimization process to determine the trade-off between cost

and environmental performances for this system. Burer et al. [5] optimize the design

and operation of a residential heating, cooling and power generation system with

3

regard to cost and CO2 emissions by a multi-criteria analysis. Even fuzzy sets have

been employed in an attempt to deal with uncertainties in multi-criteria evaluations:

Mamlook et al. [6], for example, use them in the comparison of solar systems.

A timely example of a residential energy supply technology with limited operational

experiences is small-scale combined heat and power generation (micro-CHP, micro-

cogeneration) within a residential building. In this paper, we analyze the

competitiveness of micro-CHP as one of ten alternative heating systems for a Finnish

single-family house. These traditional systems are represented by district heating,

geothermal heating, electrical heating applying baseboards, floor heating or auxiliary

heat sources (fireplace, solar heating, air heat pump), oil heating (with or without

solar heat collectors), and traditional natural gas heating. A micro-CHP heating

system is defined as a traditional natural gas heating system equipped with a 1 kWe

Solid-Oxide Fuel Cell and floor heating.

In our analysis, we pay attention to the uncertainties that pertain to financial and

environmental performance; we also account for incomplete information about the

decision-maker’s preferences by using the PAIRS method (Preference Assessment by

Imprecise Ratio Statements) [7] which is a multi-criteria decision making (MCDM)

methodology based on multi-attribute value theory (MAVT) (see, e.g.,

Keeney&Raiffa [8]). In the MAVT framework, the problem is structured as a value

tree where the topmost attribute stands for overall decision objectives and where the

attributes on the lower levels are used to measure to what extent these objectives are

attained. The relative importance of attributes is expressed through non-negative

weights that are normalized so that they add up to one.

4

In the present context, one of the advantages of PAIRS is that it admits incomplete

information through intervals and enables several modes of sensitivity analysis. As a

result, PAIRS is suitable for examining how the competitiveness of a micro-CHP

heating system depends on alternative assumptions about the decision makers’

preferences, for example. While PAIRS and related methods (e.g., PRIME,

Preference Ratios in Multiattribute Evaluation) [9], have been extensively applied in

other contexts (e.g. Gustafsson et al. [10] estimate the market capitalization value of a

new technology-based company), the evaluation of residential energy supply systems

is a new field of application.

In the remainder of this paper, we first review results from the performance

evaluation of residential heating systems and characterize relevant parameters

through respective intervals. We then outline the PAIRS methodology, motivate its

use in this evaluation problem, and present results from the comparison of ten heating

systems for a Finnish single-family house.

2 The sources of uncertainty

2.1 Energy use

In the evaluation of residential energy supply systems, the error of energy demand is

mainly related to simulation errors and user-specific reasons, such as the use of

electrical appliances [11]. There are also other sources of error, such as the difference

between statistical and measured electricity demand. Experiences suggest that the

variation between actual energy demand and energy demand estimated on the basis of

statistics or simulations is at most ± 10 % [12,13].

5

2.2 Techno-economic parameters

Data on techno-economic parameters is often acquired from i) the literature, ii)

statistical data, and iii) expert judgments. The temporal (long-term) economic

uncertainty in life cycle cost evaluations is mainly associated with the selection of

discount rate, the prices of technology, electricity and fuels, the life span of

technology and the characteristics of technology, such as efficiency. In liberalized

energy markets, real estate owners may choose their energy supplier, which results in

variability in end-user energy prices. At the moment, there is no consensus about the

buyback price of electricity (i.e., the monetary compensation which an electricity

producer receives per kWh of electricity he feeds into the grid). Actually, it may vary

between zero and the full retail price of electricity.

2.3 Use of materials

The amount of materials can be determined if the exact composition of each product

is known in the entire system. Because this information is rarely available, an

estimate is needed. When evaluating environmental burdens due to the use of natural

resources, one must consider resources contained by the end product as well as

resources that are consumed by the production process; the latter constitute the

“ecological rucksack” of the product. The total material input can be assessed

employing the material input factor (MI) that indicates the amount of abiotic and

biotic materials, water and air that is invested in producing a kilogram of a certain

material or a kilowatt-hour of energy. Material input factors depend on several issues

related to the production chains and technology and can vary significantly. A good

example is copper whose abiotic material input factor varies from 179 kg kg-1 (the

world average in 2003) to 500 kg kg-1 (produced in Germany in 1998); hence, the

uncertainties in material input factors should be recognized [14].

6

2.4 Emissions

Carbon dioxide equivalent (CO2 equivalent) and sulphur dioxide equivalent (SO2

equivalent) indicate global warming and acidification, expressing the amount of

carbon or sulphur dioxide that causes the same global warming or acidification as

some amount of another gas. For example, according to the International Panel on

Climate Change (IPCC), one kilogram of methane had a global warming potential

equal to 23 kg of carbon dioxide in 2001. The equivalents – which depend on

technology and production chains – can be determined for any material and form of

energy, provided that the above relations (e.g. 23 kgCO2/kgmethane) for various gases

are known. For example, when light heating oil is burnt in a boiler the efficiency of

which is 93 %, the amount of released CO2 and methane is 0.284 kg and 0.004 g per

one kilowatt of fuel, respectively [15]. The CO2 equivalent is 0.284 + 0.001·23·0.004

≈ 0.284 kg CO2 per one kilowatt of fuel.

2.5 Conclusions of the uncertainty

Table 1 presents estimates for the ranges of above parameters, based on the above

statements and the earlier review by Alanne et al. [11]. In effect, life-cycle costs are

minimized when the energy demand and price are at their lower bound, when the

discount rate is at its upper bound, the unit price of a micro-CHP plant as well as

service and maintenance costs are at their lower bound, and when the buyback price

of electricity, investment support and overall efficiency are at their upper bounds.

Environmental burdens, in turn, are minimized when the life span of a system is as

long as possible, and the use of material and emissions attain their lower bounds.

(Insert Table 1 around here.)

7

3 Methodology

3.1 General frame

If a decision is based solely on the minimization of costs, the decision maker’s

problem is simple (i.e., select the alternative with the lowest cost). However, the

problem becomes more challenging when environmental burdens should be

minimized at the same time when costs, too, are to be minimized. This is because the

minimization of costs and environmental burdens are usually contradictory

objectives, as it is often expensive to utilize environmentally friendly products.

In consequence, the selection of the heating system of a building is inherently a

multi-criteria problem where both financial and environmental issues must be

accounted for. This problem can be approached with value tree analysis that has a

solid foundation in multi-attribute value theory (MAVT) [8]. In MAVT, the decision

problem is structured as a tree (see: Figure 1) where the relevant objectives are

modelled by corresponding attributes. Attributes on the higher level correspond to

general objectives (e.g., minimization of life-cycle costs). These attributes are

decomposed into more specific attributes on the lower levels such that numerical (or

otherwise unambiguous) measurement scales can be attached to the twig-level

attributes at the lowest level of the value tree.

(Insert Figure 1 around here.)

8

3.2 The overall value of a heating system

Applications of MAVT [22] are often based on additive models where low

performance with regard to some attribute can be compensated by high performance

with regard to another (e.g., high life-cycle costs can be compensated by low

environmental burdens). In the additive model, the overall value (see: main goal in

Fig.1) of a heating system is the weighted sum of its attribute-specific scores with

regard to the relevant attributes, i.e.,

1

N

i ii

S w=

= ∑ s (1)

where S is the overall value of a heating system, N is the number of twig-level

attributes, wi is the normalized weight (between 0 and 1) of the i-th attribute, and si is

the normalized single-attribute score associated with the achievement level of a

heating system on the i-th attribute.

The relative importance of the i-th attribute is expressed in terms of its normalized

weight wi in the unit interval [0,1]. At the two extremes, the case wi=0 corresponds to

the case where the i-th attribute is irrelevant, while wi=1 means that the evaluation is

based on the i-th attribute only.

Scores are typically normalized onto the [0,1] range. If the performance of the

heating system with regard to the i-attribute is to be maximized and approximated by

a linear value function, the normalized score si can be calculated from

,min

,max ,min

i ii

i i

a as

a a−

=− (2)

where ai is the achievement level of a heating system with respect to the i-th attribute,

ai,min is the lowest achievement level of all the alternative heating systems with

9

respect to the i-th attribute, and ai,max is the highest achievement level of all the

alternative heating systems with respect to the i-th attribute. Conversely, when the i-

th attribute is to be minimized in the same way, the normalized score si is obtained

from

,max

,max ,min

ii

i i

a as

a a−

=−

i (3)

The following example illustrates the determination of overall value for a micro-CHP

heating system using the value tree in Fig. 1. For the time being, we assume that the

attributes are equally important in the sense that they have equal weights; explicit

preference statements will be introduced later, i.e., the following computations are

merely illustrative.

The weights of the higher level attributes (0.33 + 0.33 + 0.33 = 1) add up to one, as

well as the weights of the twig-level attributes (0.167 + 0.167 + 0.167 + 0.167 + 0.33

= 1).

The alternatives consist of ten different heating systems, referred to as System 1

through 10, with micro-CHP heating as the last one. The corresponding achievement

levels are shown in Table 2.

(Insert Table 2 around here.)

In terms of life-cycle costs, System 4 is the best alternative because its life-cycle

costs are lowest (35,550 EUR), while System 8 has the highest costs (65,900 EUR).

If the decision-maker’s (DM) preferences for life-cycle costs are linear, the score for

10

life-cycle costs for a micro-CHP heating system (System 10) can thus be obtained

from Eq. (3) as

65900 65250 0.02.65900 35550LCC

s −= =−

Normalized scores and criteria weights for a micro-CHP heating system are presented

in Table 3.

(Insert Table 3 around here.)

Based on the above weight and score information, the overall value of a micro-CHP

heating can thus be obtained from Eq. (1) as

0.33 0.02 0.167 1.00 0.167 1.00 0.167 0.52 0.167 0.71 0.55.CHPS = ⋅ + ⋅ + ⋅ + ⋅ + ⋅ =

3.3 Application of PAIRS

PAIRS method (Preference Assessment by Imprecise Ratio Statements) [7]) is a

multi-criteria decision making (MCDM) methodology where problem structuring is

carried out in accordance with the usual principles of value tree analysis (see, e.g.,

[8]). In PAIRS, uncertainties about scores and weights are captured through interval-

valued parameters. Based on these parameters, the corresponding lower and upper

bounds for the alternatives´ overall values (value intervals) in Eq. (1) can be

computed with linear programming.

More specifically, we consider m alternative heating systems (j=1,…,m) and n

attributes (i=1,…,n). The lower bounds for normalized scores are denoted by smin,11…

smin,ji… smin,mn and the upper bounds smax,11… smax,ji… smax,mn. In PAIRS, information

about the attribute weights is elicited through pairwise comparisons where the DM

11

specifies lower and upper bounds for the relative importance between two attributes

at a time. More specifically, such a comparison yields the linear constraints

iij ij

j

wlw

≤ ≤ u

ji

(4)

where lij and uij are the lower and upper bounds that for the ratio between the weights

of the i-th and j-th attributes, respectively.

The overall value interval can be calculated from linear programs (LP)

min, max,1 1

min ,maxn n

j i ji ii i

S w s w s= =

∈

∑ ∑ (5)

where the minimization and maximization problems are solved subject to constraints

that are imposed on the attribute weights (i.e., non-negativity constraints wi ≥ 0 ,

normalization constraints , and preference statements in (4)).

i∀

11

n

ii

w=

=∑

The relative superiority of one alternative to another can be determined through

dominance structures and decision rules. In PAIRS, alternative A is better than

alternative B in the sense of absolute dominance, if the least possible value (cf.

benefit) of A is greater than the largest possible value of B; in this case the value

intervals of the two alternatives do not overlap. If the value intervals overlap, it may

be possible to conclude that one alternative is better than another on the basis of

pairwise dominance: specifically, A dominates B in the sense of pairwise dominance

if and only if the overall value of A exceeds that of B for all combinations of feasible

score and weight parameters.

When dominance relationships do not hold, decision recommendations can be

provided by using decision rules (see [8]). The rules are (i) maximax (choose the

12

alternative with the highest possible overall value), (ii) maximin (choose the

alternative for which the lowest possible value is highest), (iii) minimax regret

(choose the alternative for which the greatest loss of value relative to some alternative

is smallest), and (iv) central values (choose the alternative for which the midpoint of

the value interval lies highest).

4 Comparative analysis of heating systems

4.1 Alternatives

In the following comparative analysis, we examine the overall performance of

following heating systems:

1. District heating with floor heating (S1)

2. Geothermal heating (with heat pump and bore hole) with floor heating (S2)

3. Electrical floor heating (S3)

4. Electric baseboards (S4)

5. Electric baseboards + fireplace (heating 1-2 times per week) (S5)

6. Electric baseboards + fireplace (heating 1-2 times per week) + solar heating +

air heat pump (S6)

7. Oil heating with floor heating (S7)

8. Solar oil heating with floor heating (S8)

9. Natural gas heating with floor heating (S9)

10. Natural gas heating with 1 kWe Solid Oxide Fuel Cell and floor heating (S10)

System S10 represents micro-CHP where a Solid Oxide Fuel Cell plant (Fuel Cell

Technologies Ltd) is added to the gas heating system. The Solid Oxide Fuel Cell

system is run at its full power throughout a year so that all heat is led to the heat

distribution and domestic hot water system via the heat storage tank. The excess heat

13

and electricity are managed using the heat storage, a heat dump valve, the gas boiler,

and an electricity service with bi-directional metering.

4.2 Life-cycle costs

The annual energy consumption and life cycle costs for Systems S1-S8 were obtained

from the Finnish Energy Agency that estimates energy demand on the basis of the

Finnish Standard D5. The energy demand of a Solid Oxide Fuel Cell heating system

was determined as in Alanne et al. [23]. The parameters in Table 1 were modeled as

intervals.

4.3 Environmental burdens

To determine the environmental burdens of Systems S1-S10, the composition of the

equipment was evaluated by consulting the suppliers of energy systems, system

designers and contractors, and the literature. The annual material input, the annual

global warming (CO2-equivalent) and the acidification (SO2-equivalent) potential per

one gross square meter of a house, including both the construction and operation of

an energy supply system were estimated employing the factor data provided by

Wuppertal Institute, the Intergovernmental Panel on Climate Change (IPCC), the

Building Information Foundation (RTS) and the Technical Research Centre of

Finland.

Material Input Per Service Unit (MIPS) method has not been previously applied to

the evaluation of the environmental burdens of a residential micro-CHP heating

system. Because the air consumption provided by the Material Input Per Service Unit

(MIPS) method overlaps with the effects of emissions on global warming, the air

input that is normally included into Material Input Per Service Unit (MIPS) method

was omitted. Moreover, “the use of biotic materials” was not significant, either.

14

All the parameters in Table 1 were modeled as intervals. The intervals for life-cycle

costs (30 yr) and environmental burdens for each system containing the accumulated

uncertainty are summarized in Table 4.

(Insert Table 4 around here.)

4.4 PAIRS model

The PAIRS decision model was built with the WinPRE©1 decision support tool. The

attributes were defined so as to capture the most significant factors on which the

“value” of a residential energy supply system depends.

The scores were set by normalizing the scores into the unit interval [0,1]. Based on

the data in Table 4, the score information summarized in Table 5 was obtained from

Eqs. (2) and (3).

(Insert Table 5 around here.)

4.5 Results

To illustrate the impact of uncertainties in both score and weight information, we

show results from an illustrative example where the preference statements put

considerable weight on the environmental point of view, based on the authors’

perceptions. The score intervals in Table 5 are applied.

1 The WinPRE© software is available free of charge for research and teaching purposes at http://www.decisionarium.hut.fi.

15

The attribute weights are characterized through intervals, where the decision maker

(DM) specifies through ratio comparisons how many times more (or less) important

one attribute is with regard to another. Fig. 2 illustrates the weight elicitation in the

present analysis as a WinPRE© screenshot. For example, the expression

“w(LifeCycl)< 0,3 w(Abiotic)” in the first row in Fig. 2 indicates that life-cycle costs

have been evaluated less than 0.3 times as important as the use of abiotic resources.

(Insert Figure 2 around here.)

The alternatives’ value intervals and dominance structures are in Fig.3 and Fig.4,

respectively.

(Insert Figure 3 around here.)

(Insert Figure 4 around here.)

The pairwise comparison in Figure 4 suggests that micro-CHP (S10) is a preferred

alternative at least to district heating (S1), electrical heating (S3-S5), oil heating (S7)

and natural gas heating without electricity generation (S9). The status of micro-CHP

(S10) as a recommended alternative is also highlighted by both the maximin and

maximax decision criteria. This can be explained by noting that the life-cycle costs of

micro-CHP are relatively high and involve considerable uncertainties. Thus, if cost

considerations are deemed less important, the impacts of changes in cost-related

achievement levels become less significant; this is a major reason for why the micro-

CHP performs so well.

16

5 Summary and conclusions

In this paper, we have formulated the selection of a residential energy supply system

as a multi-criteria decision-making problem where both life-cycle costs and

environmental impacts are addressed (e.g., global warming, acidification,

consumption of natural resources). Specifically, we have examined the

competitiveness of a micro-CHP heating system with traditional heating alternatives

for a single-family house in Finland, with a particular emphasis on the impact of

several sources of uncertainties. The analyses have been carried out with the PAIRS

methodology by introducing preference statements that place considerable weight on

the minimization of environmental burdens.

The numerical results suggest that micro-CHP heating can be a viable alternative,

although no definitive conclusions about the overall value of different energy supply

options can be made due to the presence of uncertainties. In practice, the overall

value of alternatives depends on decision-makers’ preferences and techno-economic

data, which in turn, are strongly affected by the location. Thus, caution is called for

when interpreting the results of this work for the purpose of giving advice in settings

where cogeneration is considered as an energy supply alternative for single-family

houses. This notwithstanding, the results suggest that PAIRS can be a useful tool in

the evaluation of residential heating systems.

This research opens up several possibilities for future work. For example, one can

combine similar multi-criteria decision analyses with forecasting models so that score

information is obtained from confidence intervals around the forecasts. Such analyses

would make it possible to assess when specific technologies are likely to outperform

others, which would give support for the development of optimal investment

17

strategies. Furthermore, it would be of interest to construct portfolio optimization

problems where the techno-economic parameters of micro-CHP are utilized to

determine the optimal project configuration of residential energy system in new

construction and renovation projects. In future work, it would also be of interest to

expand the set of attributes (for example, safety and reliability issues were not

covered here).

Acknowledgements

The funding of Fortum Foundation for this research is gratefully acknowledged.

18

References

[1] K. Alanne, A. Saari, Sustainable Small-scale CHP Technologies for Buildings:

The Basis for Multi-perspective Decision-making, Renewable & Sustainable Energy

Reviews 8(5) (2004) 401-431.

[2] M. Riihimäki, K. Mikkola, The Willingness of a Builder of a Detached House to

Ecological Ways of Building (in Finnish) Report, Finland: VTT Building and

Transport, Tampere, 2001.

[3] P. Huovila, M. Sunikka, J. Leinonen, J. Nieminen, MCDM Using EcoProP,QFD

and MCDM-23, IEA Task 23, Report, Finland: VTT Building Technology, Espoo,

1999.

[4] H. Li, M. Burer, Z. Song, D. Favrat, F. Marechal, Green Heating System:

Characteristics and Illustration with Multi-Criteria Optimization of an Integrated

Energy System, Energy 29(2) (2004) 225-244.

[5] M. Burer, K. Tanaka, K. Yamada, D. Favrat, Multi-Criteria Optimization of a

District Cogeneration Plant Integrating a Solid Oxide Fuel Cell–Gas Turbine

Combined Cycle, Heat Pumps and Chillers, Energy 28(6) (2003) 497-518.

[6] R. Mamlook, B.A.Akash, S. Nijmeh, Fuzzy Sets Programming to Perform

Evaluation of Solar Systems in Jordan, Energy Conversion and Management 42(14)

(2001) 1717-1726.

[7] A.A. Salo, R.P. Hämäläinen, Preference Assessment by Imprecise Ratio

Statements, Operations Research 40 (1992) 1053-1061.

[8] R.L. Keeney, H. Raiffa, Decisions with Multiple Objectives: Preferences and

Value Trade-Offs, John Wiley & Sons, New York, 1976.

[9] A. Salo, R.P. Hämäläinen, Preference Ratios in Multiattribute Evaluation

(PRIME) - Elicitation and Decision Procedures under Incomplete Information, IEEE

Transactions on Systems, Man, and Cybernetics, 31/6 (2001) 533-545.

[10] J. Gustafsson, A. Salo, T. Gustafsson, PRIME Decisions©: An Interactive Tool

for Value Tree Analysis, in: M. Köksalan, S. Zionts (eds.), Multiple Criteria Decision

Making in the New Millennium, Lecture Notes in Economics and Mathematical

Systems 507, Springer-Verlag, Berlin, 2001, 165-176.

[11] K. Alanne, A. Saari, A. Salo, Comparative analysis of the life-cycle costs of

residential energy supply technologies, Energy and Buildings. [Manuscript]

[12] P. Tuomaala, Implementation and Evaluation of Air Flow and Heat Transfer

Routines for Building Simulation Tools, Doctoral thesis (Helsinki University of

19

Technology), VTT Publications 471, VTT Research Centre of Finland, Espoo, 2002.

ISBN 951-38-5995-9

[13] B. Fuehrlein, S. Chandra, D. Beal, D. Parker, R. Vieira, Evaluation of

EnergyGauge® USA, A Residential Energy Design Software, Against Monitored

Data, in: Proceedings of ACEEE 2000 Summer Study, American Council for an

Energy Efficient Economy, Washington, DC, August 2000, pp 2.115 - 2.126.

[14] M. Ritthoff, H. Rohn, C. Liedtke, Calculating MIPS. Material Productivity of

Products and Services, Wuppertal Spezial 27e, Wuppertal: Wuppertal Institute for

Climate, Environment and Energy at the Science Centre North Rhine-Westphalia,

2002, 52 p.

[15] Talotekniikan LCA-laskentaohjelman käsikirja. Technical Research Centre of

Finland. 28th March, 2003 [in Finnish]. Available at:

http://80.81.172.117/files/files/lcamanuaali2003.doc

[16] H. Manczyk, Life Cycle Cost Analysis: Selection of Heating Equipment,

Manczyk Energy Consulting, 2003. Available at:

http://www.energy.rochester.edu/efficiency/Life_cycle_cost_analysis.pdf

[17] T. Collins, S.A. Parker, V. Baxter, Assessment of Hybrid Geothermal Heat

Pump Systems, Technology Installation Review, The U.S. Department of Energy,

2001.

[18] M.W. Ellis, Fuel Cells for Building Applications, ASHRAE, USA, 2002.

[19] H.I. Onovwiona, V.I. Ugursal, Residential Cogeneration Systems: Review of the

Current Technology, Renewable and Sustainable Energy Reviews 10 (5) (2004) 1-43.

[20] L. Vihermaa, M. Lettenmeier, A. Saari, Luonnonvarojen kulutus Suomen

rautatieliikenteessä (The consumption of natural resources in Finnish railroad traffic),

Liikenne- ja viestintäministeriön julkaisuja 56/2005 [in Finnish].

[21] P. Sinivuori, A. Saari, MIPS Analysis of Natural Resource Consumption in Two

University Buildings, Building and Environment 41(5) (2006) 657-668.

[22] D.L. Keefer, G.W. Kirkwood, J.L. Corner, Perspective on Decision Analysis

Applications, 1990–2001, Decision Analysis 1(1) (2004) 4-22.

[23] K. Alanne, A. Saari, V.I. Ugursal, J. Good, The financial viability of an SOFC

cogeneration system in single-family dwellings, Journal of Power Sources 158(1)

(2006) 403-416.

http://80.81.172.117/files/files/lcamanuaali2003.dochttp://www.energy.rochester.edu/efficiency/Life_cycle_cost_analysis.pdf

20

The use of abiotic resources kg/gross-m2,a

The use of water kg/gross-m2,a

SO2-equivalent kg/gross-m2,a

CO2-equivalent kg/gross-m2,a

MIN The use of abiotic resources (0.167)

MIN The use of water (0.167)

MIN Acidification Potential (0.167)

MIN Global Warming Potential (0.167)

MAX The value of a heating system (1.00)

MIN Environmental Burden due to Natural Resource Consumption (0.33)

MIN Environmental Burden due to Emissions (0.33)

MIN Life-Cycle Costs (0.33)

Life-cycle costsEUR

Figure 1. A value tree with higher-level and twig-level attributes.

21

Figure 2. Weight elicitation in the numerical example.

22

Figure 3. Value intervals for incomplete score and preference information.

23

Figure 4. Pairwise dominance for incomplete preference information.

24

Table 1. Confidence intervals for various parameters.

Parameter MIN MAX Energy Use Electricity demand, error-% A -10 +10 Primary energy demand, error-% A -10 +10 Economic parameters Real interest rate, % B 2 6 Price of district heat, error-% C -5 +5 Price of natural gas, error-% C -5 +5 Price of electricity, error-% D -10 +10 Price of oil, error-% E -10 +10 The buyback price of electricity, %F 0 100 Investment support, %G 0 50 The unit price of a micro-CHP plant, EUR H 5,000 8,000 Micro-CHP service costs, EUR a-1 0 160I Micro-CHP maintenance costs, EUR a-1 J 200 500 Technological parameters Micro-CHP total efficiency, % K

Life span error, a L 75 -5

85 5

Material use Material use, error-% M Abiotic material input factor, error-% N Biotic material input factor, error-% N Material input factor of water, error-% N Material input factor of air, error-% N

-5 -20 0

-30 -15

+5 +20

0 +30 +15

Emissions Global Warming Potential, error-% O Acidification Potential, error-% O

-10 -10

+10 +10

A an estimate on the basis of Tuomaala [12], Fuehrlein et al. [13] B an estimate based on market interest rates and the works of e.g. Manczyk [16], Collins et al.

[17] C the maximum difference between the average price and the price given by a selected company D estimated on the basis of data provided by the Finnish Energy Market Authority E an estimate based on the uncertainty related to other energy prices F expressed as the ratio of the buyback price and retail price of electricity G an estimated percentage of the capital costs of a micro-CHP plant. H estimated on the basis of price list provided by www.fuelcellstore.com I the estimate of Finnish Energy Agency for the annual service cost of a heat conversion

system in a Finnish single-family house. J If a micro-CHP plant is substituted by a new one 1-2 times during the time period of 30 years,

and the capital cost of a micro-CHP plant is 5000 - 8000 EUR kWhe-1, the annual maintenance costs vary between 200-500 EUR a-1. The interest rate, however, is not accounted for in this estimate.

K an estimate based on Ellis [18] and Onovwiona et al. [19] L an estimate based on the interview of a Finnish life-cycle specialist M an estimate based on the composition given by a Finnish trader of energy equipment and

reference estimate presented by the Technical Research Centre of Finland (VTT) N an estimate based on three sources (Vihermaa et al. [20], Sinivuori et al.[21], and Wuppertal

Institute) O an estimate based on [15], the boiler efficiency is assumed to vary between 85 % and 95 %

http://www.fuelcellstore.com/

25

Table 2. Source data for the comparative analysis.

System

Life-cycle Costs (EUR) (30 yr)

The use of abiotic resources (kg m-2 a-1)

The use of water (ton m-2 a-1)

Global Warming Potential (kg CO2 m-2 a-1)

Acidification Potential (kg SO2 m-2 a-1)

1 47,550 81 7.54 33 0.11 2 56,700 49 15.83 22 0.07 3 36,500 64 22.59 30 0.10 4 35,550 61 22.00 29 0.10 5 38,250 59 20.79 28 0.09 6 41,200 42 14.39 19 0.06 7 60,850 45 9.14 40 0.07 8 65,900 40 7.65 35 0.06 9 55,600 43 9.05 37 0.10 10 65,250 25 0.05 29 0.07

26

Table 3. Scores and weights of a micro-CHP heating system in the example.

Attribute a -> si wi

Life-cycle Costs (30 a), EUR 65,250 -> 0.02 0.330 The use of abiotic resources, kg m-2 a-1 25 -> 1.00 0.167 The use of water, t m-2 a-1 0.05 -> 1.00 0.167 Global Warming Potential, kg m-2 a-1 29 -> 0.52 0.167 Acidification Potential, kg m-2 a-1 0.07 -> 0.71 0.167

27

Table 4. Intervals for life-cycle costs and environmental burdens.

System

Life-cycle costs (30 a) (EUR)

Abiotic (kg m-2 a-1)

Water (ton m-2 a-1)

Global warming (kg CO2 -m-2 a-1)

Acidification (kg SO2 -m-2 a-1)

S1 37,300...57,800 73...89 7…8 27…40 0.088…0.131 S2 46,000…67,400 45...54 14…17 17…26 0.058…0.086 S3 24,900…48,100 58...70 20…25 25…37 0.081…0.121 S4 23,300...47,800 55...67 20…24 24…36 0.078…0.117 S5 26,600…49,900 53...65 19…23 23…34 0.074…0.111 S6 32,300…50,100 38…47 13…16 16…24 0.051…0.077 S7 45,800…75,900 41…49 8…10 32…48 0.057…0.082 S8 52,300…79,500 36…44 7…8 28…42 0.050…0.071 S9 43,400…67,800 39…47 8…10 30…44 0.084…0.125 S10(CHP) 44,600…85,900 23…27 0…0 23…35 0.060…0.091 Lowest 23,300 23 0 16 0.050 Highest 85,900 89 25 48 0.131

28

Table 5. Normalized score intervals for life-cycle costs and environmental burdens.

System

Life-cycle costs (30 a) Abiotic Water Global warming Acidification

S1 0.45…0.78 0.00…0.24 0.67…0.73 0.26…0.66 0.00…0.53 S2 0.30…0.64 0.52…0.67 0.30…0.43 0.68…0.95 0.55…0.90 S3 0.60…0.97 0.29…0.48 0.00…0.18 0.36…0.73 0.12…0.61 S4 0.61…1.00 0.33…0.52 0.03…0.20 0.39…0.75 0.18…0.65 S5 0.58…0.95 0.36…0.54 0.08…0.25 0.44…0.79 0.25…0.70 S6 0.57…0.86 0.63…0.77 0.36…0.48 0.76…1.00 0.67…0.98 S7 0.16…0.64 0.60…0.73 0.60…0.67 0.00…0.49 0.60…0.90 S8 0.10…0.54 0.68…0.80 0.66…0.72 0.19…0.61 0.74…1.00 S9 0.29…0.68 0.64…0.76 0.60…0.67 0.12…0.57 0.08…0.58 S10 (CHP) 0.00…0.66 0.93…1.00 1.00…1.00 0.41…0.77 0.50…0.87

1 Introduction2 The sources of uncertainty2.1 Energy use2.2 Techno-economic parameters2.3 Use of materials2.4 Emissions2.5 Conclusions of the uncertainty

3 Methodology3.1 General frame3.2 The overall value of a heating system3.3 Application of PAIRS

4 Comparative analysis of heating systems4.1 Alternatives4.2 Life-cycle costs4.3 Environmental burdens4.4 PAIRS model4.5 Results

5 Summary and conclusionsAcknowledgementsReferencesAttribute